Computing Power for Training

Languages at Hugging Face
1

@patrickvonplaten
Thank you for your great work!
I am wondering how much computing power to replicate the pre-trained XLSR-base model. (In terms of number of GPUs or TPUs)